Full-scale parameter optimization of biologically constrained biophysical computational model of the hippocampal network

Ivan Soltesz, Stanford University

Usage Details

The computing resources on the Blue Waters supercomputer will support our study of the mechanisms of hippocampal information processing. They will enable the rigorous and detailed computational characterization of brain function related to spatial navigation and episodic memory as well as a broad computational exploration of which network parameters are most important for optimal brain function.

With previous allocations on Blue Waters, we have completed our work on constructing and tuning the CA1 model; substantially updated the previously developed dentate gyrus (DG) network model; and made initial progress towards building a full-scale CA3 model. Upon reaching the CA3 model construction milestone, we will begin the process of connecting the three model networks.

Drawing on a BRAIN initiative grant, the project team previously studied how the hippocampal formation generates sharp-wave-ripples (SWRs), which are thought to represent replay of episodic memory sequences and are required for subsequent memory recall; as part of this effort, we are constructing the first full-scale model of the CA3 region of the hippocampus, which will be connected with the existing DG and CA1 models to provide insight into the dynamical properties of hippocampal networks that produce the feature selectivity and specific oscillatory patterns in select neuronal ensembles.  In turn, these ensembles encode location information and generate SWR events. Drawing on NIH and NASA grants, the team used its full-scale models to model the mechanisms of abnormal dynamics that emerge in epilepsy or through radiation damage.